NATURE GENETICS| VOLUME 39| NUMBER 3| MARCH 2007 281 a more accurate depiction of the actual situation (Fig. 4 in ref. 2). Third, we have already shown that in the normal population, there are identical STOX1 variations (including the Y153H variant that affects an absolutely conserved amino acid in winged helix proteins) 2. These variations can be even more frequent in specific populations (such as the Han Chinese in Beijing; http://www. hapmap. org). However, it is unknown whether these variations segregate maternally and within the same genetic context in nonaffected females or in females affected with nonfamilial or sporadic forms of preeclampsia. Information on this is lacking in the data from Moore and co-workers1. Most importantly, in their families, the existence of a parent-oforigin effect (for example, by comparison of the haplotypes shared between mother and affected daughters) has not been explored4. We predict that this effect is absent in nonfamilial and maternal forms of preeclampsia. With respect to the imprinting status of the CTNNA3 gene, an approach that includes information on cell identity (with discrimination between villous and extravillous trophoblast) is mandatory to assess the imprinting status of CTNNA3 (ref. 5). This parallels the situation seen for CDKN1C5. Without this, no definitive conclusion can be reached. We do agree that the imprinting status of